A counterpulsation control method and apparatus

Abstract

The application discloses a counterpulsation control method and device, and belongs to the technical field of medical instruments. The application aims to solve the technical problem of the clinical safety risk of the existing counterpulsation control technology. The counterpulsation control method comprises the following steps: extracting a quality feature from an electrocardiosignal, judging whether the electrocardio quality is qualified; when the electrocardio quality is qualified, calculating an electrocardio abnormality degree of the electrocardiosignal; determining whether a heart beat timing parameter can be used for counterpulsation control based on the quality feature, an individual electrocardio reference and the electrocardio abnormality degree; judging whether a predicted takeover qualification is valid based on a heart beat timing parameter prediction model and the electrocardio abnormality degree; generating a counterpulsation timing by using the heart beat timing parameter and / or the heart beat timing parameter prediction model according to whether the heart beat timing parameter can be used for counterpulsation control and whether the predicted takeover qualification is valid; and generating a counterpulsation control instruction according to the counterpulsation timing. The application can avoid the clinical safety risk that the counterpulsation is directly stopped when the electrocardiosignal is temporarily unidentifiable, and the patient loses assistance.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, specifically to a counterpulsation control method and device. Background Technology

[0002] Severe cardiac diseases such as heart failure, acute myocardial infarction, and cardiogenic shock can severely impair the heart's pumping function, leading to decreased left ventricular stroke volume, reduced coronary perfusion pressure, and inadequate perfusion of systemic organs. In these situations, drug therapy alone is often insufficient to maintain effective circulation, necessitating the use of mechanical circulatory support devices to provide additional hemodynamic support, maintain perfusion of vital organs, and create conditions for the recovery of cardiac function.

[0003] Intraatrial balloon pump (IABP) and percutaneous ventricular assist device (pVAD) are commonly used mechanical circulatory support methods in clinical practice. IABP involves placing a balloon in the aorta, which inflates during diastole to increase coronary perfusion pressure and deflates during presystole to reduce left ventricular ejection resistance. A left ventricular pVAD pumps blood from the left ventricle into the aorta via a catheter, directly unloading the left ventricle. IABP relies on the R-wave timing in the electrocardiogram (ECG) to determine the deflation trigger time and the T-wave timing to determine the inflation trigger time, achieving synchronization between balloon inflation / deflation and cardiac systolic / diastolic rhythm. PVAD, on the other hand, relies on the R-wave and T-wave timing to determine the cardiac cycle and synchronously adjust the pump's operation.

[0004] In the ICU environment, factors such as unstable electrode adhesion, body movement, sweat, lead wire traction, electromyographic activity, power frequency interference, and defibrillation can all cause sampling interruption, baseline drift, or waveform distortion, resulting in the R / T timing becoming unidentifiable for a short period of time. However, the assistive device already inserted into the patient's body does not lose its clinical assistance needs due to the temporary deterioration of the signal.

[0005] However, existing IABP and pVAD control technologies combine ECG display quality, R / T detection results, and counterpulsation control availability into a single process. When ECG signals are interfered with, causing the R / T timing to become temporarily unrecognizable, existing technologies directly equate this temporary unrecognizable signal with the condition of stopping counterpulsation. This causes the assist device already installed in the patient to discontinue its support, resulting in the patient losing diastolic coronary perfusion support and systolic load reduction, leading to a discontinuity in hemodynamic status before and after assist, posing a significant clinical safety risk. Summary of the Invention

[0006] The present invention primarily addresses the technical problem of clinical safety risks associated with existing counterpulsation control technologies.

[0007] According to the first aspect, one embodiment provides a counterpulsation control method, comprising:

[0008] Real-time acquisition of electrocardiogram signals;

[0009] Quality features are extracted from the electrocardiogram (ECG) signal, and ECG quality is calculated based on the quality features to determine whether the ECG quality is qualified.

[0010] When the ECG quality is acceptable, heartbeat timing parameters are extracted from the ECG signal, and the degree of ECG abnormality of the ECG signal is calculated based on the individual ECG reference and the heartbeat timing parameters.

[0011] Based on the quality characteristics, individual ECG reference, and degree of ECG abnormality, it is determined whether the cardiac timing parameters can be used for counterpulsation control;

[0012] Based on a pre-established cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the predicted takeover qualification is valid;

[0013] Based on whether the cardiac timing parameters can be used for counterpulsation control and whether the predicted takeover qualification is valid, the counterpulsation timing is generated using the cardiac timing parameters and / or the cardiac timing parameter prediction model.

[0014] Counterpulsation control commands are generated based on the counterpulsation timing.

[0015] In one embodiment, the quality characteristics reflect at least one of the following: sampling integrity of the ECG signal, baseline stability, noise level, degree of saturation or clipping, R / T waveform discernibility, and waveform morphology consistency.

[0016] In one embodiment, calculating the degree of electrocardiographic abnormality of the electrocardiogram signal based on an individual electrocardiogram reference and the heartbeat timing parameters includes:

[0017] The individual ECG reference is updated based on the cardiac timing parameters and the ECG characteristics of the ECG signal;

[0018] The heartbeat timing parameters and the electrocardiogram characteristics of the electrocardiogram signal are compared with the updated individual electrocardiogram reference, and the deviation of the heartbeat timing parameters and the electrocardiogram characteristics relative to the individual electrocardiogram reference is calculated.

[0019] The degree of ECG abnormality is determined based on the deviation.

[0020] In one embodiment, the cardiac timing parameters include at least one of R wave timing, T wave timing, adjacent cardiac RR interval, and intracardiac RT interval;

[0021] Determining whether the cardiac timing parameters can be used for counterpulsation control includes: determining whether the R wave timing can be used for counterpulsation control, and determining whether the T wave timing can be used for counterpulsation control.

[0022] In one embodiment, determining whether the cardiac timing parameters can be used for counterpulsation control based on the quality characteristics, individual ECG reference, and degree of ECG abnormality includes:

[0023] Calculate the R-wave control error budget and the T-wave control error budget;

[0024] The cardiac timing parameters are compared with the individual ECG reference, and the location reliability of the R wave timing and the location reliability of the T wave timing are determined by combining the quality characteristics and the degree of ECG abnormality.

[0025] Based on the R-wave control error budget and the positioning reliability of the R-wave timing, determine whether the R-wave timing can be used for counterpulsation control;

[0026] Based on the T-wave control error budget and the positioning reliability of the T-wave timing, determine whether the T-wave timing can be used for counterpulsation control.

[0027] In one embodiment, the R-wave control error budget and the T-wave control error budget are calculated in the following manner:

[0028] The R-wave control error budget is calculated based on the allowable phase deviation of the venting target, the venting path execution time, and the venting reserve margin.

[0029] The T-wave control error budget is calculated based on the allowable phase deviation of the inflation target, the inflation path execution time, and the inflation margin.

[0030] And / or,

[0031] Whether the R-wave timing and T-wave timing can be used for counterpulsation control is determined by the following methods:

[0032] When the R-wave control error budget is greater than 0 and the positioning reliability of the R-wave timing is greater than the first reliability threshold, it is determined that the R-wave timing can be used for counterpulsation control.

[0033] When the T-wave control error budget is greater than 0 and the positioning reliability of the T-wave timing is greater than the second reliability threshold, it is determined that the T-wave timing can be used for counterpulsation control.

[0034] In one embodiment, the step of generating a counterpulsation timing using the heart rate timing parameters and / or a heart rate timing parameter prediction model based on whether the heart rate timing parameters can be used for counterpulsation control and whether the predicted takeover eligibility is valid includes:

[0035] When both the R-wave timing and the T-wave timing can be used for counterpulsation control, a counterpulsation timing is generated based on the R-wave timing and the T-wave timing.

[0036] When the R-wave timing is available for counterpulsation control but the T-wave timing is not, the latest decision time for the T-wave is determined. If the target T-wave is obtained before the latest decision time for the T-wave and the target T-wave timing is available for counterpulsation control, the counterpulsation timing is generated based on the R-wave timing and the target T-wave timing. If the target T-wave is not obtained before the latest decision time for the T-wave, the predicted RT interval is generated based on the R-wave timing using the cardiac timing parameter prediction model, and the counterpulsation timing is generated based on the R-wave timing and the predicted RT interval.

[0037] When the R-wave timing is unavailable for counterpulsation control and the predictive takeover qualification is valid, based on the last known ECG state before the R-wave timing becomes unavailable, the predicted R-wave timing value and the predicted RT interval value are generated through the cardiac timing parameter prediction model, and the counterpulsation timing is generated based on the predicted R-wave timing value and the predicted RT interval value.

[0038] In one embodiment, the method further includes:

[0039] During periods when the R-wave timing is unavailable for counterpulsation control, the validity of the predicted takeover eligibility is reassessed; if the predicted takeover eligibility is deemed invalid, counterpulsation timing generation is stopped; and / or,

[0040] When the R-wave timing is restored to be usable for counterpulsation control, the first R-wave timing available for counterpulsation control is used as the starting point. A counterpulsation timing is generated based on this starting point, and the cumulative time between the R-wave timings unusable for counterpulsation control and the starting point is not included in the adjacent heartbeat RR interval. Unexecuted counterpulsation timings that conflict with the starting point are cancelled. And / or,

[0041] During the period when the R-wave timing is unavailable for counterpulsation control, the last reliable ECG state before the R-wave timing became unavailable is saved; when the R-wave timing is unavailable for counterpulsation control, the R-wave timing prediction value and the RT interval prediction value are prevented from updating the individual ECG reference and the cardiac timing parameter prediction model.

[0042] In one embodiment, determining the validity of the predicted takeover eligibility based on a pre-established cardiac timing parameter prediction model and the degree of ECG abnormality includes:

[0043] For historical heart rate timing parameters that are determined to be usable for counterpulsation control, the predicted values ​​of subsequent heart rate timing parameters are calculated using the heart rate timing parameter prediction model.

[0044] The predicted value is compared with the corresponding actual heartbeat time series parameters, and the prediction error of the heartbeat time series parameter prediction model is calculated.

[0045] The prediction confidence is calculated based on at least one of the prediction error, the rhythm stability of the historical heartbeat timing parameters, the degree of ECG abnormality, and the counterpulsation execution margin.

[0046] When the prediction confidence level is greater than the third confidence level threshold, the prediction takeover qualification is deemed valid.

[0047] When the prediction confidence level is less than or equal to the third confidence level threshold, the prediction takeover qualification is deemed invalid.

[0048] According to a second aspect, one embodiment provides a counterpulsation control device, comprising:

[0049] Memory, used to store programs;

[0050] A processor for implementing the steps in the method involved in the first aspect by executing the program.

[0051] According to a third aspect, one embodiment provides a counterpulsation control device, comprising:

[0052] ECG acquisition module, used to acquire ECG signals in real time;

[0053] The electrocardiogram (ECG) quality assessment module is used to extract quality features from the ECG signal, calculate the ECG quality based on the quality features, and determine whether the ECG quality is qualified.

[0054] The timing extraction module is used to extract heartbeat timing parameters from the electrocardiogram signal when the electrocardiogram quality is qualified.

[0055] An ECG abnormality evaluation module is used to calculate the degree of ECG abnormality of the ECG signal based on an individual ECG reference and the heartbeat timing parameters when the ECG quality is qualified.

[0056] The counterpulsation availability determination module is used to determine whether the cardiac timing parameters can be used for counterpulsation control based on the quality characteristics, individual ECG reference, and degree of ECG abnormality.

[0057] The prediction module is used to pre-establish a cardiac timing parameter prediction model, and based on the cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the prediction takeover qualification is valid.

[0058] The predictive takeover scheduling module is used to generate a counterpulsation timing based on whether the cardiac timing parameters are available for counterpulsation control and whether the predictive takeover qualification is valid, using the cardiac timing parameters and / or a cardiac timing parameter prediction model.

[0059] A counterpulsation control command generation module is used to generate counterpulsation control commands according to the counterpulsation timing sequence;

[0060] The counterpulsation execution module is used to execute the counterpulsation control commands.

[0061] The counterpulsation control method according to the above embodiments extracts quality features and calculates ECG quality, using the qualification of ECG quality as a prerequisite for extracting cardiac timing parameters and calculating the degree of ECG abnormality. Then, the quality features, individual ECG reference, and degree of ECG abnormality are jointly mapped to an independent judgment result of whether the cardiac timing parameters can be used for counterpulsation control. This establishes an interpretable judgment relationship from signal quality to control availability, overcoming the defect of directly equating ECG display quality or R / T detection results with counterpulsation control availability. When cardiac timing parameters are unavailable, the takeover qualification is verified by combining a prediction model with the degree of ECG abnormality. Counterpulsation is maintained using the predicted timing, avoiding the clinical safety risk of directly stopping counterpulsation when the signal is briefly unrecognizable, causing the patient to lose assistance. This achieves simultaneous protection of counterpulsation continuity and control safety. Attached Figure Description

[0062] Figure 1 A flowchart of one embodiment of the counterpulsation control method;

[0063] Figure 2 A flowchart illustrating the calculation of the degree of electrocardiographic abnormality in an electrocardiogram signal, according to one embodiment;

[0064] Figure 3 A flowchart illustrating one embodiment of determining whether cardiac timing parameters can be used for counterpulsation control;

[0065] Figure 4 This is a flowchart illustrating one embodiment of determining whether a pre-selection takeover qualification is valid;

[0066] Figure 5 A flowchart illustrating the generation of counterpulsation timing in one embodiment;

[0067] Figure 6 This is a first component block diagram of a counterpulsation control device according to one embodiment;

[0068] Figure 7 This is a block diagram of the components of an electrocardiogram abnormality evaluation module according to one embodiment;

[0069] Figure 8 This is a block diagram of the counterpulsation availability determination module according to one embodiment;

[0070] Figure 9 This is a block diagram of the composition of a prediction module according to one embodiment;

[0071] Figure 10 This is a block diagram of the predictive takeover scheduling module in one embodiment;

[0072] Figure 11 This is a second component block diagram of an embodiment of a counterpulsation control device. Detailed Implementation

[0073] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0074] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0075] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this invention, unless otherwise specified, include both direct and indirect connections (linkages).

[0076] After a pVAD catheter or intra-aortic balloon catheter is inserted into a patient, the patient is usually in a stage of treatment where circulatory function is unstable, coronary perfusion is insufficient, or left ventricular afterload needs to be reduced. At this time, the continuity of counterpulsation is clinically valuable. A temporary deterioration in the electrocardiogram monitoring signal does not mean that the patient no longer needs counterpulsation, nor does it mean that the inserted auxiliary catheter can stop functioning.

[0077] If the pVAD or IABP host uses only the detection of R or T waves in the current cycle as the condition for continuing counterpulsation, momentary loosening of electrode contacts, patient turning over, or brief electrical interference can all cause interruptions in the inflation / deflation sequence. Sudden interruption of counterpulsation may reduce diastolic aortic pressure gain and coronary perfusion support, weaken systolic deburring effects, and cause discontinuities in hemodynamic status before and after support. Therefore, it is necessary to maintain counterpulsation with reliable timing evidence even when ECG is temporarily unavailable, rather than directly equating "temporarily unrecognizable signal" with "treatment no longer needed."

[0078] Therefore, this invention provides a counterpulsation control method. The overall concept of this invention is as follows: the counterpulsation host continuously acquires ECG data and calculates quality characteristics within a predetermined time window or on a per-beat basis to calculate the ECG quality. High-quality real heartbeats are used to establish the patient's own RR, RT, and morphological references; waveforms with unqualified quality must not update this reference. The host further determines whether the current rhythm is abnormal and whether the predicted pattern still holds based on the degree of deviation of the current ECG from the patient's reference. Based on this, the host evaluates the positioning reliability of the real R and real T, respectively, and compares them with the counterpulsation phase allowance error, valve and pneumatic path execution time, and reserved margin to obtain the usable state of the R timing and the usable state of the T timing.

[0079] When both true R and T are available, the prediction is controlled by the true ECG. When true R is available but T is unavailable, true R is retained and bounded substitution is applied to RT. When true R is unavailable, the prediction timing is allowed to maintain counterpulsation only if the predictor (predictive model) has been validated by previous true heartbeats and the current prediction still has execution margin. The prediction model undergoes learning and shadow validation during normal ECG, with predicted values ​​compared to the true values ​​of the next beat on a beat-by-beat basis. After loss of calibration, the last reliable state is frozen, and the prediction timing only participates in the control of the current cycle and does not participate in model updates. After the true ECG recovers, the first reliable true R that has been recovered is used as the new anchor point, and the new true RR is calculated from this anchor point, without considering the accumulated time of loss of calibration as a cardiac cycle.

[0080] In this embodiment of the invention, "available" does not mean that the ECG waveform can be displayed or that a certain detection peak exists. Rather, it means that the R-sequence, T-sequence, or corresponding cardiac cycle parameters obtained from the ECG waveform have a positioning error and remaining execution time that meet the counterpulsation timing requirements for balloon deflation and inflation. "Predicted" refers to the substitution of temporarily missing timing within the same ECG counterpulsation control chain and does not involve switching to arterial pressure, pacing, or other trigger sources.

[0081] In one embodiment of the present invention, a counterpulsation control method 1000 is provided, such as... Figure 1 As shown, the method includes:

[0082] Step 110: Real-time acquisition of electrocardiogram (ECG) signals.

[0083] In some embodiments, electrocardiogram (ECG) signals can be continuously acquired via lead electrodes attached to the patient's body surface. The acquisition leads can be single-lead or multi-lead, and the sampling frequency can be 250Hz, 500Hz, or 1000Hz, selected based on the device model and pneumatic response time requirements. Simultaneously with ECG signal acquisition, sampling status information is recorded, including whether there is a sampling interruption within the current ECG sampling window, whether the sampling points are continuous, and whether there is a saturation cutoff where the amplitude reaches the upper limit of the acquisition device's range. This information is used in subsequent quality feature extraction steps.

[0084] It is understood that the acquisition of ECG signals is not limited to surface leads, but can also be achieved using intracavitary electrodes or other methods; the sampling frequency and number of leads can be flexibly configured according to the equipment model, pneumatic characteristics and clinical application scenarios, and this application does not limit specific sampling parameters.

[0085] Step 120: Extract quality features from the electrocardiogram (ECG) signal, calculate the ECG quality based on the quality features, and determine whether the ECG quality is up to standard.

[0086] In some embodiments, the quality characteristics reflect at least one of the following: sampling integrity of the ECG signal, baseline stability, noise level, degree of saturation or clipping, R / T waveform discernibility, and waveform morphology consistency.

[0087] As an optional implementation method, the electrocardiogram signals acquired within the current k-th electrocardiogram time window are... Extracting quality features , It can include the following components: sampling integrity, i.e., the proportion of valid sampling points in the current ECG time window to the total sampling points; baseline stability, i.e., the degree of deviation of the baseline drift amplitude of the current ECG time window from the mean of the recent window baseline; noise level, i.e., the ratio of high-frequency noise energy to signal energy; saturation or clipping ratio, i.e., the proportion of sampling points whose amplitude reaches the upper limit of the acquisition range; R / T waveform identifiability, i.e., the significance and location reliability of the peak values ​​of candidate R waves and T waves in the ECG signal; and recent waveform consistency, i.e., the correlation between the current candidate R / T waveform and the ECG template.

[0088] Among them, the ECG template is a predefined universal standard ECG waveform template, which includes the typical waveform morphology characteristics of normal R waves and T waves and their corresponding characteristic reference ranges. The characteristic reference range can cover the allowable range of various quality characteristic components under normal ECG acquisition conditions, such as sampling integrity, baseline stability, noise level, saturation or clipping ratio, R / T waveform identifiability, and recent waveform morphology consistency. It is suitable for ECG quality evaluation of various types of patients.

[0089] Quality characteristics It can be extracted in the following ways:

[0090] (1)

[0091] in, Let be the quality feature extraction function, representing the extraction from the ECG time window. The mapping function for extracting quality features can be implemented using statistical calculations, template correlation analysis, etc. The specific implementation method can be determined according to the actual situation, and this invention does not impose any restrictions.

[0092] Extract quality features Afterwards, ECG quality It can be represented as:

[0093] (2)

[0094] in, For based on The inverse inference mapping function of the function, through the quality feature vector Perform reverse inference to obtain the current ECG time window. Corresponding overall ECG quality This reflects the overall level of current ECG acquisition quality; The specific implementation can be a rule, a piecewise function, a statistical model, or a learning model; this application does not limit the specific implementation method.

[0095] As an optional implementation method, ECG quality can be... Compare with the corresponding feature reference range in the ECG template, when When all components fall within the characteristic reference range defined by the ECG template, the ECG quality is deemed acceptable, allowing the current real heartbeat timing parameters to be written into the individual ECG reference, and proceeding to subsequent steps of heartbeat timing parameter extraction and ECG abnormality degree calculation; when If any component deviates from the characteristic reference range defined by the ECG template, the ECG quality is deemed unqualified, and the data of the current ECG time window is prevented from updating the individual ECG reference. The candidate peak of the current window is not used as a reliable real heartbeat in subsequent processing.

[0096] In the ICU environment, ECG acquisition may be affected by a variety of factors, such as unstable electrode adhesion, body movement, sweat, lead wire traction, electromyographic activity, and power frequency interference. These factors can cause sampling interruption, baseline drift, random noise, saturation clipping, R / T waveform distortion, or peak position shift.

[0097] On the other hand, premature beats, sudden changes in heart rate, irregular rhythms, altered R / T intervals, or changes in electrocardiographic morphology can also disrupt the original cardiac cycle pattern. Poor acquisition quality and true rhythm abnormalities have different causes, but both can lead to unidentifiable R / T timing or render the next cycle unsuitable for extrapolation. Therefore, this embodiment of the invention first assesses the acquisition quality, then uses qualified true heartbeats to establish a patient-specific reference, and subsequently evaluates the degree of abnormality and counterpulsation availability.

[0098] Step 130: When the ECG quality is acceptable, extract the heartbeat timing parameters from the ECG signal, and calculate the degree of ECG abnormality based on the individual ECG reference and the heartbeat timing parameters.

[0099] In this embodiment of the invention, the cardiac timing parameters include at least one of the following: R-wave timing, T-wave timing, RR interval between adjacent cardiac beats, and RT interval within a cardiac beat. Individual ECG reference (patient reference) is a statistical reference reflecting the patient's own ECG patterns, formed by accumulating historical, high-quality, real cardiac beats. It may include RR reference, RT reference, and morphological reference, respectively reflecting the statistical patterns of the patient's recent cardiac cycle length, R-wave to T-wave interval, and ECG waveform morphology.

[0100] In some embodiments, such as Figure 2 As shown, based on individual ECG references and heartbeat timing parameters, the degree of ECG abnormality in the ECG signal is calculated, including:

[0101] Step 131: Update the individual ECG reference based on the ECG timing parameters and ECG signal characteristics;

[0102] Step 132: Compare the cardiac timing parameters and ECG characteristics of the ECG signal with the updated individual ECG reference, and calculate the deviation of the cardiac timing parameters and ECG characteristics from the individual ECG reference;

[0103] Step 133: Determine the current degree of ECG abnormality based on the deviation amount.

[0104] In some embodiments, the cardiac timing parameters used in steps 131 and 132 may include the adjacent cardiac RR interval and the intracardiac RT interval. The adjacent cardiac RR interval can reflect the current real cardiac cycle. The electrocardiographic features of the electrocardiogram signal include waveform morphology and temporal continuity.

[0105] In this embodiment of the invention, ECG quality assessment precedes ECG abnormality assessment. If ECG quality is insufficient, it means that the ECG signal waveform in the current window cannot reliably reflect the patient's true rhythm. Waveform abnormalities caused by acquisition artifacts or noise should not be directly identified as patient arrhythmia.

[0106] Individual ECG references are maintained and updated only by real ECGs of acceptable quality. When ECG quality is unacceptable, the timing parameters and ECG characteristics of the current window are prevented from being written to the individual ECG reference. The update of the individual ECG reference in the (k+1)th heartbeat cycle can be represented as follows:

[0107] (3)

[0108] (4)

[0109] The meanings of the parameters in the above formula are as follows:

[0110] Individual ECG reference for the kth beat;

[0111] Individual ECG reference for the (k+1)th beat;

[0112] The measured feature vector of the k-th real electrocardiogram includes at least the RR interval, RT interval, R / T wave morphological features, and temporal continuity features.

[0113] : Reference feature mapping function, used to obtain individual ECG references Extract or generate with Reference feature vectors corresponding to the dimensions;

[0114] State correction gain, used to control the measured feature vector of the current ECG. For individual electrocardiogram reference The correction range.

[0115] As a gating parameter, it takes the value of 0 or 1, which can ensure that the predicted time series and substandard data do not change the true reference of the patient.

[0116] Updating an individual ECG reference can specifically involve writing the RR interval, RT interval, and waveform morphology information of the current ECG signal into the individual ECG reference, and then using sliding window statistics, exponential recursion, or robust statistics to update the RR reference, RT reference, and morphology reference, so that the individual ECG reference continuously reflects the patient's recent true ECG pattern.

[0117] In this embodiment of the invention, an individual electrocardiogram reference (patient reference) is generated using a qualified real heartbeat. Assess the degree of current ECG abnormality. In other words, compare the current ECG signal's heartbeat timing parameters and ECG characteristics with an individual ECG reference, and calculate the following deviations: This refers to the deviation of the current RR interval from the RR reference in the individual's electrocardiogram reference. That is, the deviation of the current RT interval from the RT reference in the individual electrocardiogram reference; That is, the deviation of the current R / T waveform morphology from the morphological reference in the individual electrocardiogram reference. This refers to the deviation of the current cardiac timing continuity from the timing continuity reference in the individual's electrocardiogram reference.

[0118] Then, based on the calculated deviations... , , and This determines the current degree of ECG abnormality. It can be represented in the following ways:

[0119] (5)

[0120] in, This is a mapping function for the degree of ECG abnormality, used to merge multiple deviations into a unified abnormality index. It can be achieved using weighted summation, weighted norm, piecewise function, fuzzy mapping, or a trained classification model.

[0121] Step 140: Based on quality characteristics, individual ECG reference, and degree of ECG abnormality, determine whether the cardiac timing parameters can be used for counterpulsation control.

[0122] In some embodiments, determining whether cardiac timing parameters can be used for counterpulsation control includes: determining whether R-wave timing can be used for counterpulsation control, and determining whether T-wave timing can be used for counterpulsation control.

[0123] In this embodiment of the invention, the purpose of judging ECG quality and ECG abnormality is not to complete a general ECG diagnosis, but to further evaluate whether the positioning error of R wave or T wave timing is within the allowable range of counterpulsation control (i.e., control error budget); this embodiment of the invention independently judges the availability of R wave timing and T wave timing.

[0124] In practical applications, existing technologies often conflate ECG display quality, R / T detection results, and counterpulsation control availability. However, the ability to display or detect peak values ​​in an ECG waveform does not necessarily indicate that the peak location error meets the inflation / deflation timing requirements. Therefore, this invention, based on ECG quality evaluation, further calculates the degree of abnormality of the current ECG relative to the patient's own reference, using acquisition quality, ECG abnormality degree, and individual ECG reference together as the criteria for determining the availability of R / T timing counterpulsation. On the one hand, ECG quality evaluation can identify interference at the acquisition level, preventing spurious data from contaminating the patient's reference; on the other hand, ECG abnormality degree evaluation can reflect the degree of deviation of the current true rhythm from the patient's recent patterns, ensuring that availability determination considers not only signal clarity but also whether the current rhythm is still suitable as a counterpulsation timing basis. Together, these two factors distinguish between the fundamentally different conditions of "waveform displayability" and "location error meeting airway timing constraints," establishing an interpretable determination relationship from signal quality to counterpulsation control availability. This overcomes the deficiency of existing technologies that directly equate ECG display quality or R / T detection results with counterpulsation control availability.

[0125] Specifically, such as Figure 3 As shown, based on quality characteristics, individual ECG reference, and the degree of ECG abnormality, the determination of whether cardiac timing parameters can be used for counterpulsation control includes:

[0126] Step 141: Calculate the R-wave control error budget and the T-wave control error budget;

[0127] Step 142: Compare the cardiac timing parameters with individual ECG references, and determine the localization reliability of the R wave timing and the T wave timing respectively, taking into account the quality characteristics and the degree of ECG abnormality.

[0128] Step 143: Based on the R-wave control error budget and the positioning reliability of the R-wave timing, determine whether the R-wave timing can be used for counterpulsation control;

[0129] Step 144: Based on the T-wave control error budget and the positioning reliability of the T-wave timing, determine whether the T-wave timing can be used for counterpulsation control.

[0130] In this embodiment of the invention, for either the R-wave timing (R-sequence) or the T-wave timing (T-sequence), the control error budget available for ECG positioning error is first calculated. The R-wave control error budget is calculated based on the allowable phase deviation of the deflation target, the deflation path execution time, and the deflation margin. The T-wave control error budget is calculated based on the allowable phase deviation of the inflation target, the inflation path execution time, and the inflation margin. In other words, the R-wave or T-wave control error budget... Calculated in the following way:

[0131] (6)

[0132] in, Indicates timing, , This refers to the allowable phase deviation for the corresponding charge / discharge targets in the R-sequence or T-sequence time series. The execution time for venting or inflating the air circuit (including the execution time of valves, air circuits, and airbags, or their uncertainties). Allow for a margin when deflating or inflating. Control error budget. This represents the maximum allowable range of ECG positioning error, assuming the inflation / deflation timing requirements are met; it controls the error budget. It typically takes 10~15ms.

[0133] For example, assuming the IABP inflation target is to trigger at the beginning of diastole after the T wave, the permissible phase deviation is... The gas path execution time is ±30ms. 15ms, with a margin of safety. The time limit is 3ms, therefore the control budget is: This means that there is an allowable margin of 12ms for ECG localization error.

[0134] After obtaining the control error budget, based on the quality characteristics Degree of ECG abnormality and individual ECG reference Calculate the confidence level of the candidate R-series or T-series positioning error falling within the corresponding control error budget, i.e., the positioning confidence level of the R-series or T-series. Calculated in the following way:

[0135] (7)

[0136] in, The R or T timing sequence detected in the electrocardiogram signal. This refers to the actual R or T timing, i.e., the moments when the R or T wave actually occurs during the patient's cardiac electrical activity. It reflects the probability that the positioning error of the current detection sequence meets the control error budget.

[0137] In some embodiments, whether the R-wave timing can be used for counterpulsation control and whether the T-wave timing can be used for counterpulsation control are determined by the following methods:

[0138] When the R-wave control error budget is greater than 0 and the positioning reliability of the R-wave timing is greater than the first reliability threshold, the R-wave timing is determined to be usable for counterpulsation control.

[0139] When the T-wave control error budget is greater than 0 and the positioning reliability of the T-wave timing is greater than the second reliability threshold, the T-wave timing is determined to be usable for counterpulsation control.

[0140] The first confidence threshold and the second confidence threshold may be equal or unequal.

[0141] In other words, determine the usable state of the R-wave and T-wave time series respectively: when Meet credibility requirements and control budget When the value is greater than zero, the timing determination can be used for counterpulsation control, and the output availability status is valid; otherwise, the timing determination cannot be used for counterpulsation control, and the output availability status is invalid. The availability of R-wave timing and T-wave timing is independent of each other. For the same heartbeat, there may be situations where R is available but T is unavailable, or both R and T are available, or neither R nor T is available.

[0142] In other words, whether R-wave timing or T-wave timing can be used for counterpulsation can be expressed as follows:

[0143] (8)

[0144] (9)

[0145] in, Indicates timing, , Indicates the availability of R-wave timing. Indicates the availability of T-wave timing; This indicates the first or second confidence threshold.

[0146] when and When both are 1, it means that both the R timing and the T timing can be used for counterpulsation control. The subsequent generation of the counterpulsation timing for this cycle uses the real R timing and the real T timing, and the counterpulsation is completely driven by the real ECG.

[0147] when For 1 and When the value is 0, the R sequence can be used for counterpulsation control, while the T sequence cannot. Subsequent counterpulsation sequences generated for this cycle retain the true R sequence as the cycle anchor, only providing bounded prediction and replacement for missing RT intervals. The true R sequence still participates in counterpulsation control, only replacing missing RT intervals with predictions, without extending to the entire cardiac cycle, thus avoiding unnecessary abandonment of the entire true cardiac cycle due to the T sequence being unavailable alone.

[0148] when When the value is 0, the actual ECG is insufficient to determine the primary counterpulsation timing, and whether to use prediction must be determined based on the following prediction reliability assessment. This embodiment of the invention translates ordinary ECG quality assessment into pVAD / IABP counterpulsation control availability.

[0149] The ability to display a waveform or detect a peak does not necessarily mean that the peak positioning error meets the inflation / deflation timing requirements. This invention incorporates quality characteristics, the degree of ECG abnormality, and individual ECG references into the reliability calculation, establishing an interpretable mapping relationship from signal quality to counterpulsation control availability. This overcomes the deficiency of directly equating ECG display quality or R / T detection results to counterpulsation control availability.

[0150] Step 150: Based on the pre-established cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the predicted takeover qualification is valid.

[0151] In some embodiments, such as Figure 4 As shown, based on a pre-established cardiac timing parameter prediction model and the degree of ECG abnormality, the validity of the predicted takeover eligibility is determined, including:

[0152] Step 151: For historical heartbeat timing parameters that are determined to be usable for counterpulsation control, calculate the predicted values ​​of subsequent heartbeat timing parameters using the heartbeat timing parameter prediction model;

[0153] Step 152: Compare the predicted values ​​with the corresponding actual heartbeat time series parameters and calculate the prediction error of the heartbeat time series parameter prediction model;

[0154] Step 153: Calculate the prediction confidence based on at least one of the following: prediction error, rhythm stability of historical cardiac timing parameters, degree of ECG abnormality, and counterpulsation execution margin.

[0155] Step 154: When the prediction confidence is greater than the third confidence threshold, the prediction takeover qualification is determined to be valid.

[0156] Step 155: When the prediction confidence is less than or equal to the third confidence threshold, the prediction takeover qualification is deemed invalid.

[0157] For historical cardiac time series parameters that are determined to be usable for counterpulsation control, the predicted values ​​of subsequent cardiac time series parameters are calculated using a cardiac time series parameter prediction model. The input historical data for the cardiac time series parameter prediction model is... It includes only reliable, real heartbeats, excluding data from substandard windows and any virtual data generated by predictive time series, thus ensuring that the predictive model is always based on the patient's true rhythm patterns. The predictive model can be implemented using at least one of the following: mean-based, weighted trend, linear state-space model, or recursive filter regression model.

[0158] In other words, during a normal ECG period, the cardiac timing parameter prediction model uses reliable (i.e., reliable and usable for counterpulsation control) RR and RT values ​​to predict the R and T timing of the next ECG time window (next beat). After the actual heartbeat arrives in the next beat, the predicted values ​​from the previous beat are compared with the actual cardiac timing parameters arriving in this beat, and the prediction error of the previous beat is calculated. , Calculated in the following way:

[0159] (10)

[0160] (11)

[0161] in, This is a prediction error metric function used to measure the difference between the true time series parameters and the predicted time series parameters; For the kth beat, a reliable and genuine heartbeat, The predicted values ​​of the time series parameters of the (k+1)th beat are based on the reliable true heartbeat time series parameters of the kth beat. The actual heartbeat timing parameters for the (k+1)th beat. The difference between the two is the prediction error. The prediction error reflects how well the cardiac timing parameter prediction model fits the current patient's rhythm; the smaller the error, the more accurate the prediction model's prediction of the current rhythm.

[0162] Then, based on at least one of the following: prediction error over multiple consecutive real heartbeats, rhythm stability of those consecutive real heartbeats (i.e., historical heartbeat timing parameters), current ECG abnormality level, and counterpulsation execution margin, a mapping function is used to... Forming prediction credibility , Calculated in the following way:

[0163] (12)

[0164] in, This is a prediction confidence mapping function used to integrate historical prediction errors, rhythm stability, current ECG abnormality, and execution margin into prediction confidence. The prediction error is based on N consecutive historical real heartbeats. For the degree of rhythm stability, Given the current degree of ECG abnormality, This refers to the counterpulsation execution margin. The smaller the prediction error, the more stable the rhythm, the lower the degree of ECG abnormality, and the more sufficient the execution margin, the higher the prediction reliability. Mapping function. It can be achieved using weighted summation, rules, or statistical models; the specific form is not limited.

[0165] The validity of the predicted takeover eligibility is determined by comparing the predicted confidence level with a third confidence level threshold. It can be represented in the following way:

[0166] (13)

[0167] in, This is the third confidence threshold; For indicator functions, when the condition inside the parentheses is true, The function takes the value 1 when the condition within the parentheses is false. The function takes the value 0.

[0168] When prediction confidence Greater than the third confidence threshold At that time, the predicted takeover was determined. Qualification is valid, that is Predicting timing can take over counterpulsation control; when the prediction reliability is high... Less than or equal to the third confidence threshold At that time, the predicted takeover qualification was deemed invalid, that is The predicted timing sequence must not take over counterpulsation control. Valid prediction takeover eligibility indicates that the prediction model has been validated beat-by-beat with continuous real heartbeats, and the current prediction results meet the conditions for taking over the counterpulsation timing sequence.

[0169] Understandably, the third credibility threshold Continuous verification number N and mapping function The specific form can be flexibly configured according to the equipment model, gas path characteristics and clinical safety requirements. This invention does not limit the specific parameter values ​​and calculation methods.

[0170] In existing technologies, prediction is often only temporarily initiated after a decline in ECG quality, lacking bypass learning and out-of-sample validation during normal ECG periods, thus failing to prove the prediction results are qualified to take over counterpulsation timing. In this invention, the cardiac timing parameter prediction model runs continuously during normal real ECG periods, with the prediction results remaining in a shadow state and not participating in the control of the actual counterpulsation timing. Each time a new real heartbeat arrives, the host compares the predicted value of the previous beat with the actual value of the current beat, calculating the prediction error beat by beat and accumulating validation evidence. Only when the prediction error, rhythm stability, ECG abnormality, and counterpulsation execution margin on multiple consecutive real heartbeats all meet the requirements is the prediction deemed valid for takeover. This mechanism ensures that the prediction model has been fully validated by real heartbeats before obtaining takeover qualification, avoiding the safety risk of temporarily activating an unverified predictor to directly take over the counterpulsation timing after a decline in ECG quality.

[0171] Step 160: Based on whether the cardiac timing parameters can be used for counterpulsation control and whether the predicted takeover qualification is valid, generate the counterpulsation timing using the cardiac timing parameters and / or the cardiac timing parameter prediction model.

[0172] In some embodiments, such as Figure 5 As shown, based on whether the cardiac timing parameters can be used for counterpulsation control and whether the predicted takeover eligibility is valid, the counterpulsation timing is generated using cardiac timing parameters and / or a cardiac timing parameter prediction model, including:

[0173] Step 161: When both the R-wave timing and the T-wave timing can be used for counterpulsation control, generate the counterpulsation timing based on the R-wave timing and the T-wave timing.

[0174] Step 162: When the R-wave timing is available for counterpulsation control but the T-wave timing is not, determine the latest decision time for the T-wave. If the target T-wave is obtained before the latest decision time for the T-wave and the target T-wave timing is available for counterpulsation control, generate the counterpulsation timing based on the R-wave timing and the target T-wave timing. If the target T-wave is not obtained before the latest decision time for the T-wave, generate the predicted RT interval based on the R-wave timing using the cardiac timing parameter prediction model, and generate the counterpulsation timing based on the R-wave timing and the predicted RT interval.

[0175] Step 163: When the R-wave timing is not available for counterpulsation control and the predicted takeover qualification is valid, based on the last known ECG state before the R-wave timing is unavailable, the predicted R-wave timing value and the predicted RT interval value are generated through the cardiac timing parameter prediction model, and the counterpulsation timing is generated based on the predicted R-wave timing value and the predicted RT interval value.

[0176] In this embodiment of the invention, the counter-pulse timing of the current cycle is selected according to the available status of the R timing sequence, the available status of the T timing sequence, and the predicted takeover qualification, in the following three cases.

[0177] Scenario 1: When both the R-sequence and T-sequence can be used for counterpulsation control, the current real R-sequence and real T-sequence are directly used to generate the counterpulsation sequence for this cycle, which is then used to drive the inflation / deflation control command. Simultaneously, the cardiac timing parameter prediction model continues to run in shadow mode, calculating the prediction error beat by beat and updating the prediction reliability, continuously accumulating verification data for potential subsequent predictive takeover.

[0178] The second scenario: When the R time series is available but the T time series is temporarily unavailable, the true R time series is retained as the anchor point for this cycle to determine the starting time of the counterpulsation time series for this cycle. Then, based on the historical qualified true RT interval and the current cycle (RR) trend, the predicted RT interval value is generated through the cardiac timing parameter prediction model. Based on the actual R-sequence, the latest decision time of the actual T-wave is calculated by subtracting the inflation execution time and inflation margin from the target inflation reference time. , Calculated in the following way:

[0179] (14)

[0180] in, For actual R timings, Inflation execution time, Allow sufficient margin for inflation.

[0181] exist Previously, when a reliable true T was obtained, the true T (target T wave timing) was determined using the true RT; otherwise, the predicted RT was used. The predicted RT interval value can be generated by the cardiac timing parameter prediction model, and the RT interval value can be represented in the following way:

[0182] (15)

[0183] in, This refers to the actual RT interval used in this cycle. To assess the reliability of the target T-wave timing sequence for localization. For the arrival time of the target T-wave; For the true RT interval, This represents the predicted RT interval.

[0184] The predicted RT interval is used only for inflation timing within the current cycle. It does not generate spurious T events, is not written to the real T event queue, and is not used to update the RT reference in the individual ECG reference.

[0185] At the latest decision moment of T-wave Previously, if the target T-wave was obtained and its T-sequence could be used for counterpulsation control, the actual R-sequence and the target T-sequence were used to generate the counterpulsation sequence for this cycle; if it reached If a usable target T wave is still not obtained, the actual R time series and the predicted RT interval are used to generate the cost cycle counterpulsation time series.

[0186] If the target T-wave arrives before the inflation control action is executed and at a time no later than [time missing], This allows for the overriding of unexecuted RT interval predictions; already executed same-cycle inflation actions retain their executed status and are not repeated. This bounded waiting mechanism limits the impact of T-sequence unavailability to the scope of local RT substitution, avoiding the unnecessary abandonment of the entire real cardiac cycle due to T-sequence unavailability alone.

[0187] The third scenario: When the true R timing is unavailable (during the period of loss of target) and the predictive takeover qualification is valid, read the last reliable ECG state snapshot before the R timing became unavailable. This snapshot may include the predictive state corresponding to the last reliable beat, the individual ECG reference, and the current counterpulsation execution status. Last reliable ECG state snapshot It can be represented in the following ways:

[0188] (16)

[0189] The meanings of the parameters in the above formula are as follows:

[0190] : The last known reliable electrical state snapshot saved before the R timing became unavailable;

[0191] The state of the model is predicted by the time series parameters of the heartbeat at the kth beat, such as historical RR, RT, trend of change, internal state of the model and prediction covariance, etc.

[0192] Individual ECG reference for the kth beat;

[0193] : The counterpulsation control phase when saving a snapshot;

[0194] : The counterpulsation execution status when the snapshot is saved, including whether inflation and deflation have been performed in this cycle and related execution flags.

[0195] The state snapshot function is used to freeze and save the aforementioned trusted state, rather than performing numerical calculations.

[0196] The purpose of this function is to provide a recovery starting point that will not be contaminated by low-quality ECG or prediction results when the actual R timing is temporarily unavailable.

[0197] Then, based on the last known electrocardiographic state before the R-wave timing becomes unavailable, the predicted R-wave timing and RT interval are generated using the cardiac timing parameter prediction model. Based on these predicted values, the counterpulsation timing for this cycle is generated to maintain the counterpulsation. The update mechanism of the cardiac timing parameter prediction model can be represented as follows:

[0198] (17)

[0199] (18)

[0200] The meanings of the parameters in the above formula are as follows:

[0201] : Predicted ECG timing status at the k-th beat;

[0202] : The predicted state for the next (k+1) beat obtained through recursion;

[0203] : State prediction function, used to predict the next state based on the current state;

[0204] The actual electrocardiographic observations at the k-th beat, such as the actual RR interval and RT interval;

[0205] : Observation mapping function, used to determine the state Generation and Real Observations Corresponding predictive observations;

[0206] The residual between actual ECG observations and predicted observations;

[0207] State correction gain is used to control the magnitude of the correction of the predicted state by the actual observation residual.

[0208] During periods when R-wave timing is unavailable for counterpulsation control, the last known reliable electrical state before the R-wave timing became unavailable is saved. During periods of loss of target, predictive timing can be generated from this last known reliable electrical state, but this can be achieved by setting... When R-wave timing cannot be used for counterpulsation control, the predicted R-wave timing and RT interval values ​​are prevented from updating the individual ECG reference and cardiac timing parameter prediction model. Even if the predicted timing cannot be written into the actual RR, RT, and individual ECG reference, the predicted timing cannot correct the model parameters. In other words, the prediction results can be used to maintain counterpulsation timing, but will not form virtual data self-training; the reverse training pathway of the predicted timing for the model is cut off, eliminating the risk of continuous prediction self-reinforcement and cumulative drift.

[0209] Understandably, the latest decision-making time for T-waves The calculation parameters (inflation execution time, inflation margin) can be flexibly determined according to the equipment model and air circuit configuration; the specific form of the prediction model is not limited to the linear state space model, but can also use the mean, weighted trend or recursive filter regression model. This invention does not limit the specific prediction model form and parameter values.

[0210] In this embodiment of the invention, the true R / T timing always takes precedence over the predicted timing, and predicted timing takeover is only allowed when the true timing is unavailable and the predicted timing is eligible for takeover. During periods of normal ECG, the cardiac timing parameter prediction model continuously executes predictions for the next cardiac timing parameters, but the prediction results remain in a shadow state. When the true R / T timing is available, the true R / T timing always takes precedence over the predicted timing, and prediction takeover is only allowed when the true R / T timing is subsequently determined to be unavailable and... Only when the forecast timing is valid can it qualify for takeover and counter-attack. The selection relationship of the forecast timing is as follows:

[0211] (19)

[0212] in, Represents the actual R / T timing. This represents the actual R-series time series and the predicted RT interval. This represents the R time series forecast and the RT interval forecast.

[0213] In addition, if the R-wave timing is not available for counterpulsation control and the predictive takeover qualification is invalid, then no counterpulsation timing will be generated and no counterpulsation will be performed.

[0214] Existing technologies use an overall validity or invalidity assessment for both R and T intervals. In actual operation, situations may arise where the R timing remains reliable while the T timing is temporarily unrecognizable. In such cases, abandoning the entire real cardiac cycle and handing over all counterpulsation control to the prediction model would unnecessarily expand the prediction range and increase control safety risks. Therefore, this application independently calculates the counterpulsation control error budget for both the R and T timings, and outputs the R and T availability states independently. When the R timing is available but the T timing is temporarily unavailable, the true R timing is retained as the anchor point for this cycle, and only the missing RT interval is replaced by a bounded prediction. The prediction range is limited to the RT region and does not extend to the entire cardiac cycle. This R / T fractional availability determination mechanism avoids the problem of unnecessarily abandoning the entire real cardiac cycle due to the T timing being unavailable alone. It minimizes the prediction control range while retaining the true R anchor point, reducing the timing deviation risk caused by whole-cycle prediction and improving the safety and continuity of counterpulsation control.

[0215] In this embodiment of the invention, during the period of loss of calibration, the predicted timing is only used for counterpulsation (inflation / deflation timing) in the current cycle, and is not written into the actual RR, RT, or individual ECG reference. The prediction results are also not used to correct model parameters. This data isolation mechanism fundamentally cuts off the reverse training pathway of the prediction timing for the prediction model, avoiding the problem of self-training and cumulative drift caused by the continuous input of virtual R, virtual T, or virtual cycles into the prediction model, so that the predicted value will not gradually deviate from the patient's true rhythm during the period of loss of calibration.

[0216] In some embodiments, the method provided by the present invention further includes:

[0217] During periods when R-wave timing cannot be used for counterpulsation control, the validity of the predictive takeover qualification is reassessed; if the predictive takeover qualification is deemed invalid, the generation of counterpulsation timing is stopped.

[0218] In this embodiment of the invention, as the loss of target continues, the current prediction span (i.e., the cumulative time length from the last reliable true R time series to the current moment) is recorded, and combined with the current counter-pulse execution margin, the prediction reliability is re-evaluated according to the prediction reliability calculation method in formula (12). It is important to note that during periods of miscalibration, due to the lack of real observations, prediction errors cannot be updated by comparing them with the true values. Reassessment of prediction reliability can only be based on the prediction span and the current execution margin. The longer the prediction span, the higher the risk of the predicted value deviating from the patient's true rhythm. Prediction reliability can only be maintained or decreased; it cannot be increased without justification due to a lack of real observations.

[0219] If the re-evaluated prediction confidence is still greater than the third confidence threshold, the prediction takeover qualification remains valid and the prediction timing continues to maintain counterpulsation; if the re-evaluated prediction confidence is less than or equal to the third confidence threshold, the prediction takeover qualification is invalid, the generation of counterpulsation timing stops, and the system enters a waiting state until the real ECG recovers and the control chain is re-established.

[0220] This mechanism prevents the predictive model from taking over counter-pulse control indefinitely after the loss of target time has been prolonged, ensuring that the predictive takeover has clear time boundaries and safe exit conditions, and avoiding the control security risks caused by long-term reliance on unverifiable predictive timing.

[0221] In some embodiments, the method provided by the present invention further includes:

[0222] When the R-wave timing is restored and can be used for counterpulsation control, the first R-wave timing that can be used for counterpulsation control is taken as the starting point, and a counterpulsation timing is generated based on the starting point. The cumulative time between the R-wave timing that cannot be used for counterpulsation control and the starting point is not included in the adjacent heartbeat RR interval, and the unexecuted counterpulsation timing that conflicts with the starting point is canceled.

[0223] In this embodiment of the invention, after the real ECG is restored, the cycle anchor point is re-established using the first real R sequence determined to be usable for counterpulsation control; that is, this real R sequence is taken as the starting point. It can be represented as:

[0224] (20)

[0225] in, This indicates the first determination of the true R timing that can be used for counterpulsation control.

[0226] Then, based on this starting point and subsequent determination of the actual R-sequence that can be used for counterpulsation control, a new RR interval is formed. , that is to The second reliable true R time series that arrives subsequently forms a new RR interval. , It can be represented in the following ways:

[0227] (twenty one)

[0228] in, This is the second reliable real R time series.

[0229] Starting from this cycle, updates to the individual ECG reference will resume, and subsequent actual R / T timing will be re-involved in the calculation of prediction error and the evaluation of prediction takeover eligibility.

[0230] Meanwhile, the last reliable true R time series before the loss of the target was... The cumulative time of uncalibrated intervals is not included in the adjacent RR intervals and is not treated as a true RR. The reason is that this cumulative time may contain multiple unobserved heartbeats. If it is treated as a single RR interval to update the individual ECG reference or prediction model, it will lead to severe distortion of the RR reference and affect the accuracy of subsequent prediction models.

[0231] While re-establishing the periodic anchor point, cancel with Unexecuted predicted timing sequences with timing conflicts are avoided to prevent conflicts between the predicted plans generated during the loss of targeting and the actual cardiac cycle after recovery. For inflation and deflation actions already performed during the loss of targeting, the executed marker is retained, and the same cycle actions are not repeated after the actual ECG recovers to avoid timing disorder caused by the repeated triggering of predicted actions and actual events in the same cycle.

[0232] Through the above mechanism, the embodiments of the present invention form a complete control chain of "ECG quality and abnormality judgment - R / T counterpulsation availability determination - verified prediction takeover - loss of standard data isolation - real ECG recovery and re-anchoring", which realizes the full-process guarantee of safely maintaining continuous counterpulsation operation when ECG signals are temporarily unavailable and smoothly reconnecting to real ECG control after recovery.

[0233] Step 170: Generate counterpulsation control commands based on the counterpulsation timing.

[0234] In some embodiments, counterpulsation control commands are then sent to the counterpulsation component for execution.

[0235] In this embodiment of the invention, control commands corresponding to the counterpulsation component type are generated according to the timing sequence determined in the counterpulsation timing sequence: for intra-aortic balloon counterpulsation devices, balloon inflation and deflation control commands are generated and sent to the pneumatic actuator through the output interface to drive the valve to complete the deflation and inflation actions at the specified time; for percutaneous ventricular assist devices, speed adjustment commands synchronized with the cardiac cycle are generated to adjust the working state of the pump according to the counterpulsation timing sequence.

[0236] While performing counterpulsation, the system maintains the markers of executed actions and records the counterpulsation actions performed in the current cycle. When the predicted timing is covered by the actual T timing, or when there is a conflict between the predicted timing and the recovered actual R timing, the actions already performed in the same cycle are not repeated to avoid triggering the counterpulsation action twice within the same cardiac cycle and to ensure the safety of counterpulsation control.

[0237] It is understood that the specific form of counterpulsation control commands can be determined according to the counterpulsation component type and device interface specifications, and this application does not limit the specific command format and output interface form.

[0238] In some embodiments, the method provided by this invention can be applied to a pVAD or IABP host, wherein the host performs continuous counterpulsation control according to the following process: the host continuously acquires electrocardiogram (ECG) signals and sampling status, extracts quality features within the current evaluation window or on a per-beat basis, and calculates ECG quality. If the current ECG quality does not meet the requirements, the patient's ECG reference is not updated using this window; if the current ECG quality meets the requirements, the true R / T, RR, and RT are extracted and compared with the patient's reference. The degree of ECG abnormality is calculated. And the reliability that the R and T positioning errors are within the counterpulsation control budget and This enables the availability of R / T fractional counterpulsation. During the actual R / T availability period, the predicted state is updated with the actual heartbeat, and the shadow state is compared with the previous prediction and the current actual time series to calculate the prediction error, stability, and takeover eligibility. If both R and T are available, use the true R / T; if R is available but T is unavailable, wait for the true T until the latest decision time, and use the predicted RT if a reliable T is not obtained; if R is unavailable and If effective, read the last reliable snapshot and use the predicted timing to maintain counterpulsation. During the loss-of-target maintenance period, the predicted timing is only used for inflation / deflation timing and is not written to the real RR, RT, or patient reference. Check if the real ECG has recovered. If it has, re-anchor using the first reliable real R, exclude the cumulative loss-of-target interval, cancel conflicting unexecuted predicted timings, and avoid repeating already executed actions. Use subsequent real R / T to re-perform shadow validation and predictive eligibility assessment before proceeding to the next control cycle.

[0239] Standard ECG quality assessment only indicates whether the waveform is clear, not whether the R / T positioning error obtained from the waveform meets the phase requirements for balloon inflation / deflation. In the method provided in this invention, the host first calculates the acquisition quality, allowing only qualified real heartbeats to be used as a patient reference; then it calculates the degree of abnormality of the current ECG relative to the patient reference; finally, it combines the R / T positioning reliability with the counterpulsation phase allowable error, airway execution time, and remaining margin, outputting the R availability status and T availability status respectively. Through this three-step decision chain, an interpretable mapping relationship is established from signal acquisition quality to counterpulsation control availability, transforming the relatively mature ECG quality detection into a counterpulsation timing availability assessment for the pVAD / IABP host. Only after identifying that the real ECG is no longer suitable for control can subsequent prediction of maintenance counterpulsation have a clear triggering condition and clinical causal relationship. This decision chain provides clear triggering conditions for subsequent predictive takeover, ensuring that the prediction of maintenance counterpulsation has a clear clinical causal basis.

[0240] If a prediction is temporarily initiated after ECG loss without being verified by the patient's current true rhythm, direct takeover may lead to erroneous inflation / deflation. The method provided in this invention utilizes real R / T learning during normal ECG and enables the prediction bypass to operate; it evaluates the prediction error of the previous beat using the real value of the next beat; and the prediction is only qualified for takeover when error, stability, abnormality level, and execution margin all meet the requirements. In this invention, real R / T always takes precedence over unexecuted prediction sequences.

[0241] The T-wave amplitude is low, making it easy to miss detection when R is still clear. If a T-wave is missed, abandoning the true R and predicting the entire cycle would expand the control range of the model. In this embodiment, when T is unavailable but R is available, the true R is retained as the cycle anchor point. The latest decision time for the T-wave is determined based on the target inflation reference time, the gas path execution time, and the reserved margin. Before the decision time, reliable true T is prioritized; only after the decision time is the predicted RT used. This forms a complete structural relationship of "true R retention—bounded T waiting—local RT substitution—true T covering before action." The predicted RT is only an internal time series; it does not fabricate T events and does not participate in model training.

[0242] In existing technologies, continuous prediction re-inputs the virtual cycle into the predictor, leading to self-reinforcement and cumulative drift. During recovery, multiple unobserved heartbeats may be merged into an excessively long RR, or the predicted action and the actual event may be executed repeatedly. This invention freezes the last reliable state when the prediction fails to be anchored, using the prediction result only for timing control; it uses real data gating to block virtual data updates; and after recovery, it re-anchors with the first reliable real R and forms a new RR from subsequent real Rs.

[0243] In summary, the beneficial effects of the method provided by the embodiments of the present invention include at least the following:

[0244] By extracting quality features and calculating ECG quality, the suitability of ECG quality is used as a prerequisite for extracting cardiac timing parameters and calculating the degree of ECG abnormality. The quality features, individual ECG reference, and the degree of ECG abnormality are then jointly mapped to an independent judgment result indicating whether cardiac timing parameters can be used for counterpulsation control. This establishes an interpretable judgment relationship from signal quality to control availability, overcoming the deficiency of directly equating ECG display quality or R / T detection results with counterpulsation control availability. When cardiac timing parameters are unavailable, the predictive model of cardiac timing parameters, combined with the degree of ECG abnormality, determines whether the predicted takeover qualification is valid. If the takeover qualification is valid, the predictive model can be used to generate counterpulsation timing to maintain counterpulsation, avoiding the clinical safety risk of directly stopping counterpulsation when the signal is briefly unrecognizable, thus depriving the patient of assistance. This achieves simultaneous protection of counterpulsation continuity and control safety.

[0245] When ECG is temporarily unavailable due to unstable electrode contact, hair, sweat, body movement, or clinical electrical interference, signal loss is not directly equated with stopping counterpulsation. This provides a timing basis for maintaining continuous support in patients already undergoing intervention. ECG quality, abnormalities, and R / T positioning errors are linked to the counterpulsation execution error budget to avoid misinterpreting normal ECG readings as controllable counterpulsation. R and T are assessed separately, ensuring reliable R while retaining the true R even when T is missed, only replacing the missing RT, reducing unnecessary whole-cycle predictions. The predictor undergoes beat-by-beat verification with real heartbeats during normal ECG, providing traceable error evidence for its takeover qualification. Predictive data during out-of-calibration periods is not used to back-train the model, reducing self-contamination and cumulative drift from continuous extrapolation. After real ECG recovery, re-anchoring is performed and cumulative long intervals are excluded, reducing cycle distortion, planning conflicts, and repetitive actions during the recovery phase. Theoretical mapping does not rely on a fixed filter, threshold, or prediction formula, and can cover different host systems, pneumatic responses, and patient heart rate conditions.

[0246] To better understand the methods provided in the embodiments of the present invention, the following exemplary description is provided.

[0247] Example 1: Assessment of ECG quality and availability during changes in ICU electrode contact

[0248] The patient has already had a pVAD or intra-aortic balloon catheter inserted and is receiving counterpulsation support. During patient sweating and lead traction during repositioning, the control unit observes increased baseline changes, a decreased effective sampling rate, and decreased correlation between local R / T waveforms and recent templates in some sampling windows. The control unit generates quality characteristics and overall quality; windows that do not meet quality requirements are not updated for patient reference. Furthermore, candidate peaks are not considered reliable true R / T values.

[0249] The host further determines whether the reliability of the current candidate R / T position falls within the counterpulsation control error budget. If the actual R still meets the requirements but T does not, proceed to Case 1; if the actual R also does not meet the requirements, proceed to Case 2 to check the predictive qualification formed before the ECG quality deteriorated. Since the above judgment is based on the counterpulsation execution error budget, the same waveform can lead to different conclusions about the usability of counterpulsation under different heart rates or different remaining execution times.

[0250] Scenario 1: R-wave is available, but T-wave is temporarily unavailable.

[0251] The current quality and positioning reliability of the actual R wave meet the requirements, and the host computer uses this actual R wave to determine the starting point of this cycle. The T wave has not yet reached the reliability condition due to low amplitude or local interference. The host computer forms a prediction based on the historical data of qualified actual R waves, the current RR trend, and historical prediction errors. And determine the latest decision time. .

[0252] like If a reliable true value T has been obtained previously, then the true true value RT is used; if it is reached... If no trustworthy T is found, then use This forms the inflation time for the current cycle. Predicted RT is not written to the true T event queue, nor is it used to update the patient's RT reference. If the true T recovers before the action is performed and is still earlier than... If the action has been performed, the unexecuted prediction RT will be covered; if the action has been performed, only the actual T will be recorded for subsequent quality evaluation, and the action in this cycle will not be repeated.

[0253] Scenario 2: Decreased ECG quality and continuous unusable R / T ratio

[0254] Before the ECG quality deteriorates, the predictor has already completed learning and shadow verification during continuous real heartbeats. The host computer forms the next beat prediction based on the real RR and RT, and calculates the prediction error after the real R / T of the next beat arrives; the prediction takeover qualification is determined based on the prediction error, dispersion, degree of anomaly, and execution margin.

[0255] When the current real R is unavailable, the host does not use candidate peaks of unknown quality, but freezes the last trusted state according to formula (16). During the hold period, predicted time series are generated from the same snapshot to maintain the counter-pulse, and the data gating is adjusted accordingly. Setting it to zero means that the predicted time series will not update RR, RT, patient reference, and model correction terms. The host computer simultaneously records the prediction span and recalculates it based on the prediction span and execution margin to maintain reliability.

[0256] Example 2: Quality is acceptable, but rhythm changes.

[0257] The waveforms acquired by the host are continuous, without obvious saturation, and the R / T morphology is discernible, meeting the overall quality requirements. However, the true RR or RT deviates continuously from the patient's recent reference values. The host calculates the degree of abnormality. The prediction error and prediction eligibility are then re-evaluated. If the true R / T still meets the counterpulsation time series budget, the true R / T continues to be used preferentially; abnormal states themselves do not force a switch in triggering methods. Prediction is only used to maintain counterpulsation when the true time series is unavailable and the prediction confidence is still greater than the threshold.

[0258] Example 3: Reanchoring after recovery from real ECG

[0259] After the electrodes are reattached or the interference disappears, the host computer first recalculates the acquisition quality and R-positioning reliability. The first reconfirmed true R is taken as... Used to terminate unexecuted forecast plans that conflict with it; the last true R before the loss of the target. The time is not considered as an RR. After the second recovered true R arrives, it forms... And update the patient reference from that actual cycle.

[0260] For actions already performed during the current cycle upon recovery, the host retains the execution marker, and the real events are only used for re-anchoring without being repeated. Subsequent real R / T events participate in the shadow error calculation again, and the predictor regains or loses its takeover qualification. This completes the closed-loop handover between real ECG, prediction hold, and real ECG.

[0261] Example 4: Model calibration and parameter determination

[0262] During the equipment development or validation phase, typical data such as normal ECG, lead loosening, baseline drift, electromyographic interference, power line interference, saturation clipping, missed R-wave detection, missed T-wave detection, heart rate variability, and rhythm irregularities are used to establish a mapping relationship between quality characteristics and R / T positioning errors. Based on the response tests of the equipment valves, pneumatic circuits, and balloon, the execution time and timing margin under different operating conditions are determined, thereby obtaining the counterpulsation control error budget.

[0263] Using a qualified real heartbeat playback predictor, the one-step prediction error, bias, dispersion, and error varying with the prediction span are statistically analyzed to determine the reliability mapping. Parameters can be stored in partitions according to device model, gas path configuration, heart rate zone, or patient status.

[0264] It is understood that the specific implementation forms of the various technical elements in the embodiments of the present invention are not limited to the following descriptions:

[0265] ECG quality characteristics are not limited to sampling integrity, baseline stability, noise level, saturation ratio, and R / T template correlation. They can also include sampling effectiveness, baseline variation, noise energy, periodic interference energy, or candidate peak significance, or a combination of these characteristics. ECG quality can be calculated using rules, lookup tables, piecewise functions, statistical models, or learning models, with the specific method chosen based on the equipment model and clinical application scenario.

[0266] The establishment and updating of individual ECG references are not limited to exponential recursion. They can also be achieved using sliding window statistics, robust statistics, state observers, or clustering templates. The common requirement is that they are updated only by real heartbeats of acceptable quality and are not contaminated by acquisition artifacts or prediction time series.

[0267] The calculation of the degree of ECG abnormality is not limited to weighted summation of deviations. Standardized distance, probability likelihood, cumulative deviation, change point detection, rhythm rules or classification models can also be used. Its role is to evaluate whether the recent true rhythm pattern is still suitable for prediction and extrapolation, rather than to replace doctors in making ECG diagnoses.

[0268] The prediction model for heartbeat time series parameters is not limited to the linear state-space model. It can also use the mean, median, weighted trend, nonlinear state-space model, recursive filtering or regression model, etc. The prediction form can be replaced, but it must be verified by shadow of qualified real heartbeats and must not be trained and updated by the predicted time series or unqualified data.

[0269] The three phases of predictive takeover—learning, shadow verification, and loss of target retention—can be implemented with different names or combined, as long as the technical relationship of building a model with real data—verification during normal periods—takeover after real unavailability—relinquishment after real recovery is maintained, this application does not limit the specific phase division method.

[0270] This invention can be implemented by software, programmable controller, digital signal processor, microcontroller, dedicated circuit or a combination of software and hardware; the evaluation window length, model order and various threshold parameters can be flexibly set according to the equipment model, airway characteristics, heart rate range and verification results, and this invention does not limit specific parameter values.

[0271] In one embodiment of the present invention, a counterpulsation control device is provided, comprising:

[0272] Memory, used to store programs;

[0273] A processor is used to execute the program to implement all or part of the processes of the counterpulsation control method provided in the above embodiments.

[0274] In one embodiment of the present invention, a counterpulsation control device 2000 is provided, such as... Figure 6 As shown, it includes:

[0275] The ECG acquisition module 210 is used to acquire ECG signals in real time and obtain single-lead or multi-lead ECG and its sampling status.

[0276] The ECG quality assessment module 220 is used to extract quality features from ECG signals, calculate ECG quality based on the quality features, and determine whether the ECG quality is qualified.

[0277] The timing extraction module 230 is used to extract cardiac timing parameters from the electrocardiogram signal when the electrocardiogram quality is qualified.

[0278] The ECG abnormality evaluation module 240 is used to calculate the degree of ECG abnormality based on individual ECG reference and heartbeat timing parameters when the ECG quality is qualified.

[0279] The counterpulsation availability determination module 250 is used to determine whether cardiac timing parameters can be used for counterpulsation control based on quality characteristics, individual ECG reference, and the degree of ECG abnormality.

[0280] The prediction module 260 is used to pre-establish a cardiac timing parameter prediction model, and based on the cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the prediction takeover qualification is valid.

[0281] The predictive takeover scheduling module 270 is used to generate counterpulsation timing based on whether the cardiac timing parameters are available for counterpulsation control and whether the predictive takeover qualification is valid, using cardiac timing parameters and / or a cardiac timing parameter prediction model.

[0282] Counterpulsation control command generation module 280 is used to generate counterpulsation control commands according to the counterpulsation timing sequence;

[0283] The counterpulsation execution module 290 is used to execute counterpulsation control commands.

[0284] In this embodiment of the invention, the counterpulsation control device 2000 may specifically be a pVAD or an IABP host.

[0285] In some embodiments, such as Figure 7 As shown, the ECG abnormality evaluation module 240 includes:

[0286] ECG reference update module 241 is used to update the individual ECG reference based on the ECG characteristics of the ECG signal and the heartbeat timing parameters;

[0287] The deviation calculation module 242 is used to compare the cardiac timing parameters and the electrocardiographic features of the electrocardiogram signal with the updated individual electrocardiogram reference, and calculate the deviation of the cardiac timing parameters and electrocardiographic features relative to the individual electrocardiogram reference.

[0288] The abnormality calculation module 243 is used to determine the current ECG abnormality level based on the deviation amount.

[0289] In some embodiments, the cardiac timing parameters include at least one of R-wave timing, T-wave timing, adjacent cardiac RR interval, and intracardiac RT interval; determining whether the cardiac timing parameters can be used for counterpulsation control includes: determining whether the R-wave timing can be used for counterpulsation control, and determining whether the T-wave timing can be used for counterpulsation control.

[0290] In some embodiments, such as Figure 8 As shown, the counterpulsation availability determination module 250 includes:

[0291] Error budget calculation module 251 is used to calculate the R-wave control error budget and the T-wave control error budget;

[0292] The location reliability determination module 252 is used to compare the cardiac timing parameters with the individual ECG reference, and determine the location reliability of the R wave timing and the T wave timing respectively by combining the quality characteristics and the degree of ECG abnormality.

[0293] R-wave availability judgment module 253 is used to determine whether the R-wave timing can be used for counterpulsation control based on the R-wave control error budget and the positioning reliability of the R-wave timing.

[0294] The T-wave availability judgment module 254 is used to determine whether the T-wave timing can be used for counterpulsation control based on the T-wave control error budget and the positioning reliability of the T-wave timing.

[0295] In some embodiments, the R-wave control error budget and the T-wave control error budget are calculated in the following manner:

[0296] The R-wave control error budget is calculated based on the allowable phase deviation of the venting target, the venting path execution time, and the venting reserve margin.

[0297] The T-wave control error budget is calculated based on the allowable phase deviation of the inflation target, the inflation path execution time, and the inflation margin.

[0298] In some embodiments, whether the R-wave timing can be used for counterpulsation control and whether the T-wave timing can be used for counterpulsation control are determined by the following methods:

[0299] When the R-wave control error budget is greater than 0 and the positioning reliability of the R-wave timing is greater than the first reliability threshold, the R-wave timing is determined to be usable for counterpulsation control.

[0300] When the T-wave control error budget is greater than 0 and the positioning reliability of the T-wave timing is greater than the second reliability threshold, the T-wave timing is determined to be usable for counterpulsation control.

[0301] In other words, the counterpulsation availability determination module 250 is used to compare the R / T positioning reliability with the counterpulsation allowable timing error and remaining execution time, and output the R availability status and T availability status respectively.

[0302] In some embodiments, such as Figure 9 As shown, the prediction module 260 includes:

[0303] The prediction value calculation module 261 is used to calculate the predicted value of subsequent heartbeat timing parameters based on the historical heartbeat timing parameters that are determined to be usable for counterpulsation control, using a heartbeat timing parameter prediction model.

[0304] The prediction error calculation module 262 is used to compare the predicted value with the corresponding real heartbeat time series parameters and calculate the prediction error of the heartbeat time series parameter prediction model.

[0305] The prediction confidence calculation module 263 is used to calculate the prediction confidence based on at least one of the following: prediction error, rhythm stability of historical cardiac timing parameters, degree of electrocardiographic abnormality, and counterpulsation execution margin.

[0306] The first takeover qualification judgment module 264 is used to determine the validity of the predicted takeover qualification when the prediction confidence is greater than the third confidence threshold.

[0307] The second takeover qualification judgment module 265 is used to determine that the predicted takeover qualification is invalid when the prediction confidence is less than or equal to the third confidence threshold.

[0308] In other words, the prediction module 260 only uses reliable real heartbeats to establish RR and RT predictions, and compares the predicted values ​​with the subsequent real values ​​when the real ECG is normal in order to calculate the prediction error.

[0309] In some embodiments, such as Figure 10 As shown, the predictive takeover scheduling module 270 includes:

[0310] The first counterpulsation timing generation module 271 is used to generate a counterpulsation timing based on the R-wave timing and T-wave timing when both the R-wave timing and T-wave timing can be used for counterpulsation control.

[0311] The second counterpulsation timing generation module 272 is used to determine the latest decision time for the T wave when the R wave timing is available for counterpulsation control but the T wave timing is not available for counterpulsation control; if the target T wave is obtained before the latest decision time for the T wave and the target T wave timing is available for counterpulsation control, the counterpulsation timing is generated based on the R wave timing and the target T wave timing; if the target T wave is not obtained before the latest decision time for the T wave, the predicted value of the RT interval is generated based on the R wave timing and the predicted value of the RT interval, and the counterpulsation timing is generated based on the R wave timing and the predicted value of the RT interval.

[0312] The third counterpulsation timing generation module 273 is used to generate R-wave timing prediction values ​​and RT interval prediction values ​​based on the last known electrocardiographic state before the R-wave timing becomes unavailable, and to generate counterpulsation timing based on the R-wave timing prediction values ​​and RT interval prediction values ​​when the R-wave timing is unavailable and the predicted takeover qualification is valid.

[0313] In other words, the predictive takeover scheduling module 270 selects the timing sequence for the current cycle based on the availability of R / T and the predictive reliability.

[0314] In some embodiments, such as Figure 10 As shown, the predictive takeover scheduling module 270 also includes:

[0315] The counterpulsation stop module 274 is used to re-evaluate the validity of the predictive takeover qualification when the R-wave timing is not available for counterpulsation control; when the predictive takeover qualification is determined to be invalid, the generation of the counterpulsation timing is stopped.

[0316] The recovery module 275 is used to generate a counterpulsation sequence based on the first R-wave sequence that can be used for counterpulsation control when the R-wave timing recovery is available for counterpulsation control. It also excludes the cumulative time between the R-wave sequence that cannot be used for counterpulsation control and the starting point from the RR interval of adjacent heartbeats and cancels unexecuted counterpulsation sequences that conflict with the starting point.

[0317] In some embodiments, such as Figure 11 As shown, the counterpulsation control device 2000 also includes:

[0318] The trusted state storage module 300 is used to save the last trusted ECG state before the R-wave timing is unavailable during the period when the R-wave timing is unavailable for counterpulsation control. When the R-wave timing is unavailable for counterpulsation control, it prevents the R-wave timing prediction value and RT interval prediction value from updating the individual ECG reference and cardiac timing parameter prediction model.

[0319] In other words, the trusted state storage module 300 saves the last trusted state before the target is lost and prevents the predicted time series from being updated in reverse to the real data model.

[0320] The counterpulsation control device provided in this embodiment of the invention is based on the same inventive concept as the counterpulsation control method provided in the above embodiment. The operations performed by each module in the device correspond one-to-one with the steps in the method. For the specific implementation of each module, please refer to the relevant description of the above method embodiment, which will not be repeated here.

[0321] In the above embodiments, implementation can be achieved, in whole or in part, by software, hardware, firmware, or any combination thereof. Furthermore, as those skilled in the art will understand, the principles herein can be reflected in a computer program product on a computer-readable storage medium pre-loaded with computer-readable program code. Any tangible, non-transitory computer-readable storage medium may be used, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CDs, DVDs, Blu-ray discs, etc.), flash memory, and / or the like. These computer program instructions can be loaded onto a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to form a machine, such that instructions executing on the computer or other programmable data processing apparatus can generate means for performing a specified function. These computer program instructions can also be stored in a computer-readable storage medium that can instruct the computer or other programmable data processing apparatus to operate in a particular manner, such that instructions stored in the computer-readable storage medium can form an article of manufacture, including means for implementing the specified function. The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to perform a series of operational steps on the computer or other programmable apparatus to produce a computer-implemented process, such that instructions executing on the computer or other programmable apparatus can provide steps for implementing the specified function.

[0322] This document describes various exemplary embodiments with reference to them. However, those skilled in the art will recognize that changes and modifications can be made to the exemplary embodiments without departing from the scope of this document. For example, various operational steps and components for performing operational steps can be implemented in different ways depending on the specific application or considering any number of cost functions associated with the operation of the system (e.g., one or more steps can be deleted, modified, or combined with other steps).

[0323] While the principles herein have been illustrated in various embodiments, numerous modifications to the structures, arrangements, proportions, elements, materials, and components, particularly suited to specific environments and operational requirements, may be used without departing from the principles and scope of this disclosure. These modifications and other alterations or alterations will be included within the scope of this document. Those skilled in the art will recognize that many changes can be made to the details of the above embodiments without departing from the fundamental principles of the invention.

Claims

1. A counterpulsation control method, characterized in that, include: Real-time acquisition of electrocardiogram signals; Quality features are extracted from the electrocardiogram (ECG) signal, and ECG quality is calculated based on the quality features to determine whether the ECG quality is qualified. When the ECG quality is acceptable, heartbeat timing parameters are extracted from the ECG signal, and the degree of ECG abnormality of the ECG signal is calculated based on the individual ECG reference and the heartbeat timing parameters. Based on the quality characteristics, individual ECG reference, and degree of ECG abnormality, it is determined whether the cardiac timing parameters can be used for counterpulsation control; Based on a pre-established cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the predicted takeover qualification is valid; Based on whether the cardiac timing parameters can be used for counterpulsation control and whether the predicted takeover qualification is valid, the counterpulsation timing is generated using the cardiac timing parameters and / or the cardiac timing parameter prediction model. Counterpulsation control commands are generated based on the counterpulsation timing.

2. The counterpulsation control method as described in claim 1, characterized in that, The quality characteristics reflect at least one of the following: sampling integrity of the ECG signal, baseline stability, noise level, degree of saturation or clipping, R / T waveform discernibility, and waveform morphology consistency.

3. The counterpulsation control method as described in claim 1, characterized in that, The calculation of the degree of electrocardiographic abnormality of the electrocardiogram signal based on individual electrocardiogram reference and the heartbeat timing parameters includes: The individual ECG reference is updated based on the cardiac timing parameters and the ECG characteristics of the ECG signal; The heartbeat timing parameters and the electrocardiogram characteristics of the electrocardiogram signal are compared with the updated individual electrocardiogram reference, and the deviation of the heartbeat timing parameters and the electrocardiogram characteristics relative to the individual electrocardiogram reference is calculated. The degree of ECG abnormality is determined based on the deviation.

4. The counterpulsation control method as described in claim 1, characterized in that, The cardiac timing parameters include at least one of the following: R wave timing, T wave timing, adjacent cardiac RR interval, and intracardiac RT interval; Determining whether the cardiac timing parameters can be used for counterpulsation control includes: determining whether the R wave timing can be used for counterpulsation control, and determining whether the T wave timing can be used for counterpulsation control.

5. The counterpulsation control method as described in claim 4, characterized in that, The determination of whether the cardiac timing parameters can be used for counterpulsation control based on the quality characteristics, individual ECG reference, and degree of ECG abnormality includes: Calculate the R-wave control error budget and the T-wave control error budget; The cardiac timing parameters are compared with the individual ECG reference, and the location reliability of the R wave timing and the location reliability of the T wave timing are determined by combining the quality characteristics and the degree of ECG abnormality. Based on the R-wave control error budget and the positioning reliability of the R-wave timing, determine whether the R-wave timing can be used for counterpulsation control; Based on the T-wave control error budget and the positioning reliability of the T-wave timing, determine whether the T-wave timing can be used for counterpulsation control.

6. The counterpulsation control method as described in claim 5, characterized in that, The R-wave control error budget and the T-wave control error budget are calculated in the following manner: The R-wave control error budget is calculated based on the allowable phase deviation of the venting target, the venting path execution time, and the venting reserve margin. The T-wave control error budget is calculated based on the allowable phase deviation of the inflation target, the inflation path execution time, and the inflation margin. And / or, Whether the R-wave timing and T-wave timing can be used for counterpulsation control is determined by the following methods: When the R-wave control error budget is greater than 0 and the positioning reliability of the R-wave timing is greater than the first reliability threshold, it is determined that the R-wave timing can be used for counterpulsation control. When the T-wave control error budget is greater than 0 and the positioning reliability of the T-wave timing is greater than the second reliability threshold, it is determined that the T-wave timing can be used for counterpulsation control.

7. The counterpulsation control method as described in claim 4, characterized in that, The step of generating a counterpulsation timing using the heart rate timing parameters and / or a heart rate timing parameter prediction model, based on whether the heart rate timing parameters are suitable for counterpulsation control and whether the predicted takeover eligibility is valid, includes: When both the R-wave timing and the T-wave timing can be used for counterpulsation control, a counterpulsation timing is generated based on the R-wave timing and the T-wave timing. When the R-wave timing is available for counterpulsation control but the T-wave timing is not, the latest decision time for the T-wave is determined. If the target T-wave is obtained before the latest decision time for the T-wave and the target T-wave timing is available for counterpulsation control, the counterpulsation timing is generated based on the R-wave timing and the target T-wave timing. If the target T-wave is not obtained before the latest decision time for the T-wave, the predicted RT interval is generated based on the R-wave timing using the cardiac timing parameter prediction model, and the counterpulsation timing is generated based on the R-wave timing and the predicted RT interval. When the R-wave timing is unavailable for counterpulsation control and the predictive takeover qualification is valid, based on the last known ECG state before the R-wave timing becomes unavailable, the predicted R-wave timing value and the predicted RT interval value are generated through the cardiac timing parameter prediction model, and the counterpulsation timing is generated based on the predicted R-wave timing value and the predicted RT interval value.

8. The counterpulsation control method as described in claim 7, characterized in that, The method further includes: During periods when the R-wave timing is unavailable for counterpulsation control, the validity of the predicted takeover eligibility is reassessed; if the predicted takeover eligibility is deemed invalid, counterpulsation timing generation is stopped; and / or, When the R-wave timing is restored to be usable for counterpulsation control, the first R-wave timing available for counterpulsation control is used as the starting point. A counterpulsation timing is generated based on the starting point, and the cumulative time between the R-wave timings that are not usable for counterpulsation control and the starting point is not included in the adjacent RR interval. Unexecuted counterpulsation timings that conflict with the starting point are cancelled; and / or, During the period when the R-wave timing is unavailable for counterpulsation control, the last reliable ECG state before the R-wave timing became unavailable is saved; when the R-wave timing is unavailable for counterpulsation control, the R-wave timing prediction value and the RT interval prediction value are prevented from updating the individual ECG reference and the cardiac timing parameter prediction model.

9. The counterpulsation control method as described in claim 1, characterized in that, The determination of whether the predicted takeover qualification is valid based on the pre-established cardiac timing parameter prediction model and the degree of ECG abnormality includes: For historical heart rate timing parameters that are determined to be usable for counterpulsation control, the predicted values ​​of subsequent heart rate timing parameters are calculated using the heart rate timing parameter prediction model. The predicted value is compared with the corresponding actual heartbeat time series parameters, and the prediction error of the heartbeat time series parameter prediction model is calculated. The prediction confidence is calculated based on at least one of the prediction error, the rhythm stability of the historical heartbeat timing parameters, the degree of ECG abnormality, and the counterpulsation execution margin. When the prediction confidence level is greater than the third confidence level threshold, the prediction takeover qualification is deemed valid. When the prediction confidence level is less than or equal to the third confidence level threshold, the prediction takeover qualification is deemed invalid.

10. A counterpulsation control device, characterized in that, include: Memory, used to store programs; A processor for implementing the method as described in any one of claims 1 to 9 by executing the program.

11. A counterpulsation control device, characterized in that, include: ECG acquisition module, used to acquire ECG signals in real time; The electrocardiogram (ECG) quality assessment module is used to extract quality features from the ECG signal, calculate the ECG quality based on the quality features, and determine whether the ECG quality is qualified. The timing extraction module is used to extract heartbeat timing parameters from the electrocardiogram signal when the electrocardiogram quality is qualified. An ECG abnormality evaluation module is used to calculate the degree of ECG abnormality of the ECG signal based on an individual ECG reference and the heartbeat timing parameters when the ECG quality is qualified. The counterpulsation availability determination module is used to determine whether the cardiac timing parameters can be used for counterpulsation control based on the quality characteristics, individual ECG reference, and degree of ECG abnormality. The prediction module is used to pre-establish a cardiac timing parameter prediction model, and based on the cardiac timing parameter prediction model and the degree of ECG abnormality, determine whether the prediction takeover qualification is valid. The predictive takeover scheduling module is used to generate a counterpulsation timing based on whether the cardiac timing parameters are available for counterpulsation control and whether the predictive takeover qualification is valid, using the cardiac timing parameters and / or a cardiac timing parameter prediction model. A counterpulsation control command generation module is used to generate counterpulsation control commands according to the counterpulsation timing sequence; The counterpulsation execution module is used to execute the counterpulsation control commands.

Images